By name: earnings-preview description: Pre-earnings scenario analysis — consensus estimates, bull/base/bear scenarios, key metrics to watch, and catalyst checklist. Triggers on "earnings preview for [ticker]", "preview Q[X] for [ticker]", "what to watch for [ticker] earnings", or "pre-earnings [ticker]".
Earnings Preview
Core question: "What should I expect, what should I watch, and how does each outcome affect my thesis?"
Adapted from Anthropic's equity-research earnings-preview skill, simplified for our single-analyst scale with thesis-tracker integration.
Workflow
Step 1: Gather Context (Script)
Collect recent financial data and thesis context from the REST API:
uv run python skills/earnings-preview/scripts/collect_preview.py {TICKER}
uv run python skills/earnings-preview/scripts/collect_preview.py {TICKER} --quarter Q1 --year 2026
This fetches:
- Last 8 quarters of financials (
GET /api/financials/{TICKER}/quarterly?quarters=8) - Latest metrics and margins (
GET /api/financials/{TICKER}/metrics) - Segment breakdown (
GET /api/financials/{TICKER}/segments) - Recent price action (
GET /api/financials/{TICKER}/prices?period=3mo) - Existing thesis data from
data/artifacts/{TICKER}/thesis/thesis.json - Latest catalyst calendar from
data/artifacts/{TICKER}/thesis/catalysts.json
Writes preview_{Q}_{YEAR}_raw.json to artifacts.
Step 2: Scenario Framework (AI)
The agent reads the raw data and produces:
Consensus & Recent Trends:
- Revenue, EPS, margin trends from last 4-8 quarters
- Sequential and year-over-year growth rates
- Management guidance from prior quarter (if available in thesis catalysts)
Bull / Base / Bear Scenarios:
| Scenario | Revenue | EPS | Key Driver | Expected Stock Reaction |
|---|---|---|---|---|
| Bull | ||||
| Base | ||||
| Bear |
For each scenario: what would need to happen operationally, and how it maps to thesis assumptions.
Step 3: Key Metrics & Catalyst Checklist (AI)
Identify the 3-5 things that will determine the stock's reaction:
- [Metric] vs. trend — why it matters for the thesis
- [Guidance item] — what continuation/change would signal
- [Narrative shift] — strategic changes that could move the stock
Step 4: Output
uv run python skills/earnings-preview/scripts/generate_preview.py {TICKER}
uv run python skills/earnings-preview/scripts/generate_preview.py {TICKER} --persist
Artifacts
Output goes to data/artifacts/{TICKER}/earnings/:
| File | Contents |
|---|---|
preview_{Q}_{YEAR}_raw.json |
Collected data for AI analysis |
preview_{Q}_{YEAR}.json |
Structured scenario framework |
preview_{Q}_{YEAR}.md |
Narrative earnings preview |
REST API Endpoints Used
| Endpoint | What we read |
|---|---|
GET /api/financials/{TICKER}/quarterly?quarters=8 |
Quarterly trend data |
GET /api/financials/{TICKER}/metrics |
Current margins and ratios |
GET /api/financials/{TICKER}/segments |
Segment breakdown |
GET /api/financials/{TICKER}/prices?period=3mo |
Recent price action |
GET /api/companies/{TICKER} |
Company details (sector, industry) |
Cross-Skill Reads
data/artifacts/{TICKER}/thesis/thesis.json— thesis assumptions to map scenarios againstdata/artifacts/{TICKER}/thesis/catalysts.json— upcoming catalysts and prior guidance
Important Notes
- Always note that estimates are based on historical trends, not consensus — we don't have sell-side consensus data
- Historical earnings reactions help calibrate expectations
- The preview should be created 3-5 days before expected earnings
- If a thesis exists, every scenario must reference its impact on thesis assumptions